Search Results (2,047)

Accurate large-scale crop mapping is fundamental to agricultural management. However, in Liaoning Province, undulating terrain and fragmented fields make fine crop classification challenging. In particular, corn and soybean have overlapping phenologies, which can lead to spectral and structural confusion in conventional optical–SAR feature spaces and limit mapping accuracy. This study proposes a fine crop mapping framework integrating optical phenotypic, microwave structural, and meteorological time-series features. To overcome the curse of dimensionality caused by high-dimensional heterogeneous data, an adaptive feature truncation mechanism based on the transition pattern of the marginal-gain curve was designed. Additionally, a pyramid multi-scale sliding window algorithm was constructed to optimize meteorological features, achieving dimensionality reduction and precise identification of phenologically sensitive windows. The results indicate that: (1) The multi-scale feature selection strategy effectively eliminates redundant variables and maximizes the inter-class discriminability of core features, significantly improving computational efficiency and classification performance. (2) High-frequency meteorological features provide key physiological constraints. Specifically, mid-May shortwave radiation, early October precipitation, and early August growing degree days constitute the core environmental–physiological features for distinguishing confused crops, helping to mitigate the spectral confusion of dryland crops. (3) Driven by the multi-source features, the Support Vector Machine (SVM) exhibits the optimal generalization robustness for processing high-dimensional structured data, yielding an overall classification accuracy of 91.80% and a Kappa coefficient of 0.8905. This framework provides a reliable methodological reference for high-precision crop monitoring in large-scale complex planting areas. Full article

To advance crop type mapping in support of global food and water security, this study compared three spectral configurations: (A) the full 60-band DLR Earth Sensing Imaging Spectrometer (DESIS) hyperspectral narrowband (HNB) dataset, (B) a 14-band subset of DESIS-derived HNBs aligned with the planned Landsat 10 (formerly Landsat Next) spectral configuration (400–1000 nm), and (C) DESIS-based simulations of Landsat 10 superspectral broadbands. The analysis was conducted in California’s Central Valley, hereafter referred to as “the Central Valley”, during the peak growing month of August. DESIS imagery from August 2021, 2022, and 2023 was used sequentially for model development, testing, and independent validation. Over these three years, DESIS provided extensive hyperspectral coverage of much of the 4 million hectares in the Central Valley’s. Analyses were performed on Google Earth Engine using two pixel-based supervised classifiers, Random Forest (RF) and Support Vector Machine (SVM), to differentiate three major crop classes: row crops, grapes and tree crops, and winter wheat/fallow/other. The highest overall accuracy (86%) was achieved using SVM in combination with either the full DESIS hyperspectral dataset or the 14 DESIS narrowbands corresponding to Landsat 10. This finding aligns with earlier studies showing a small number of strategically positioned narrowbands can be optimal for crop type classification. Use of the narrowband datasets resulted in substantially higher accuracy (overall accuracy of 86%) compared to the simulated Landsat 10 broadbands (overall accuracy of 75%), supporting previous studies highlighting the utility of narrowbands. Despite the high accuracy using August imagery, the study indicates more granular crop type classification will require multi-temporal observations spanning the full phenological cycle (June–October), especially for a large number of crop classes. Acquiring task-based hyperspectral imagery over such large areas throughout the growing season remains operationally challenging. In contrast, Landsat 10 superspectral imagery could provide routine coverage across seasons and years that is practical and scalable for future large area crop type mapping and agricultural monitoring. Full article

Background/Objectives: To characterize the zero-shot diagnostic behavior of three commercial multimodal large language models (MLLMs) on cardiopulmonary chest radiograph findings and to assess sex-stratified performance differences. Methods: GPT-5.4, Claude Opus 4.5, and Gemini 2.5 Pro were evaluated in 4500 pathology-specific radiograph evaluations based on frontal chest radiographs from the publicly available CheXpert dataset. Three balanced cohorts of 1500 images each were constructed for cardiomegaly, pulmonary edema, and pleural effusion (375 per sex-by-label subgroup). All models received identical zero-shot prompts requesting binary classification. The primary outcome was area under the receiver operating characteristic curve (AUC-ROC) with 95% bootstrap confidence intervals. Secondary outcomes were sensitivity and specificity. Results: A total of 4500 pathology-specific radiograph evaluations were performed across the three cohorts (2250 male and 2250 female cohort entries; mean age 58.4 ± 18.0 years). GPT-5.4 achieved the highest discrimination (AUC-ROC 0.836–0.883) but showed very low sensitivity (0.043–0.424) with near-perfect specificity (0.977–0.997). Claude Opus 4.5 showed moderate discrimination (AUC-ROC 0.698–0.761) with balanced sensitivity (0.396–0.876) and specificity (0.461–0.863). Gemini 2.5 Pro showed moderate discrimination (AUC-ROC 0.745–0.770) but favored sensitivity (0.673–0.973) at the expense of specificity (0.241–0.804). Sex-stratified analyses showed consistently higher AUC point estimates in male patients for cardiomegaly and pulmonary edema, but smaller and less directional differences for pleural effusion. Conclusions: Commercial MLLMs differ considerably in operating profiles, ranging from ultraconservative to aggressive detection, so that strong aggregate discrimination can mask sensitivity too low for reliable detection. None of the evaluated models are currently suitable for autonomous chest radiograph interpretation. Sex-stratified differences were modest but non-uniform, supporting subgroup-aware reporting rather than reliance on pooled metrics alone. Full article

The Brazilian Cerrado, a global biodiversity hotspot, is under increasing pressure from agricultural expansion and native vegetation conversion, underscoring the need for efficient monitoring to support conservation and environmental policies. In heterogeneous landscapes, land use and land cover (LULC) mapping using supervised classification methods faces a major bottleneck: the need for extensive and high-quality training datasets. To address this challenge, we propose a semi-automated, clustering-based methodology for mapping secondary vegetation within previously deforested areas, reducing training-sample requirements and enabling scalable mapping through the clustering of satellite image time series. In the first stage, an unsupervised process integrates graphics processing unit (GPU)-accelerated Self-Organizing Maps and hierarchical clustering with Dynamic Time Warping to produce spectro-temporal clusters. In the second stage, specialists label and refine these clusters by visual interpretation, transferring expert knowledge from individual pixels to grouped spectro-temporal patterns. Applied to 692,000 km² of previously deforested land in the Cerrado biome, the methodology produced a mapped secondary vegetation area of 81,209 km² (11.74%). The design-based estimated area was 98,683 ± 10,071 km², with an overall accuracy of 96.45 ± 1.52%, a user’s accuracy of 96.27 ± 2.40%, a producer’s accuracy of 79.22 ± 7.94%, and an F1-score of 86.90%. The initial cluster labeling accounted for 86.3% of the final secondary vegetation area and limited the interpretation task to approximately 3000 cluster-level decisions. Implemented in the TerraClass Cerrado 2024 cycle, the workflow reduced the secondary vegetation mapping phase from approximately two years to six months while maintaining the thematic accuracy required for large-scale operational monitoring. Full article

French coastal systems are characterized by strong environmental gradients and increasing anthropogenic pressures, resulting in rapid land cover transformations across coastal landscapes. This study investigates land cover dynamics along the northern, western, and southern French coasts using Sentinel-2 summer image time series acquired between 2015 and 2025. The research aims to identify the most dynamic coastal regions and determine where land cover transitions are most pronounced. A harmonized workflow was developed in GRASS GIS for preprocessing Sentinel imagery, generating seasonal composites, classifying land cover using a Random Forest (RF) supervised algorithm, and detecting changes through time. All imagery was processed using CORINE Land Cover (Level 1) classification nomenclature and projected to Lambert-93 (EPSG:2154). Comparative analyses were performed among the three coastal regions using statistical indicators of change intensity, persistence, and transition rates. The results reveal substantial regional differences in coastal dynamics, with the southern Mediterranean coast exhibiting the highest transformation rate (22.9% of total area changed, at 2.29% ${\mathrm{yr}}^{-1}$), followed by the northern English Channel coast (18.6%; 1.86% ${\mathrm{yr}}^{-1}$) and the western Atlantic coast (14.2%; 1.42% ${\mathrm{yr}}^{-1}$). Urbanization and natural vegetation loss were identified as dominant transition types across all regions. The study demonstrates the effectiveness of Sentinel-2 time series and open-source GRASS GIS methods for long-term coastal monitoring and provides a reproducible framework for large-scale assessments of coastal land cover dynamics in Europe. Full article

Saline–alkaline land serves as a potential arable land reserve for augmenting agricultural productivity and safeguarding food security. However, long-term monitoring of saline–alkaline land conversion remains challenging because of vegetation recovery, surface changes, hydrological modification, and agricultural phenology. Compared with CCDC and LandTrendr, the proposed MK-based framework detects conversion occurrence and timing while reducing dependence on dense observations, parameter tuning, and annual classification. This study examines the spatiotemporal dynamics of saline–alkaline land converted into paddies in Da’an City, utilizing Landsat time-series data (2007–2021) from the Google Earth Engine (GEE) platform. The analysis employed Mann–Kendall (MK) trend and mutation tests to monitor conversion processes and analyze spatiotemporal dynamics. Point-biserial correlation analysis was applied to evaluate the sensitivity of various remote sensing indices in detecting land conversion. The top fifteen indices, including the Land Surface Water Index (LSWI), Salinity Index 4 (SI4), and Salinity Index 5 (SI5), demonstrated strong correlations (|r| = 0.788–0.885) and significant pre- and post-conversion spectral differences (p < 0.01). Validation via confusion matrix confirmed that the June SI5 index attained the highest detection accuracy (overall accuracy: 94.15%; Kappa coefficient: 0.86), supporting the MK trend test’s efficacy in monitoring conversion processes. The MK mutation test achieved 80.36% temporal accuracy in determining conversion timing. The spatiotemporal analyses identified heterogeneity in saline–alkaline land conversion patterns. Spatially, large contiguous paddy fields dominated the eastern region, whereas fragmented conversion characterized the west, with minimal activity in the central zone. Temporally, the conversion area expanded rapidly before 2015 and then gradually declined, reaching a cumulative converted area of 276.29 km² by 2021. This study elucidates spatiotemporal conversion dynamics to guide sustainable land use. Full article

Leptospirosis is an important reproductive disease in bovine hosts, yet its epidemiology is still largely inferred from serology, which provides limited resolution in bovines. Although genetic studies based on genotyping and whole-genome sequencing approaches are increasing, a comprehensive overview is still lacking. We analyzed bovine-origin sequences and genome metadata from databases to characterize patterns of Leptospira species and serogroups and to describe genotyping methods and molecular markers. Metadata were retrieved from GenBank and the Institut Pasteur cgMLST databases until January 2026. Extracted variables included species, genotyping approach/markers, serological classification of isolates, sample type and origin (renal vs. genital), clinical signs, and geographic location; climates were assigned using Köppen–Geiger classification. After selection, 569 records were retrieved: 411 single-locus sequences (eight molecular markers), 95 MLST profiles, and 63 genomes, from 35 countries, with most reports from South America (57.6%). Records spanned tropical, temperate, steppe, Mediterranean, and continental-cold climates. Nine species and 14 serogroups were identified; tropical and temperate areas showed greater diversity. The main agents, L. interrogans, L. borgpetersenii, and serogroup Sejroe, were predominant and widespread. Notably, L. noguchii, L. santarosai, L. venezuelensis, and L. wolffii were mostly concentrated in the Americas. Most records were from renal samples (68.7%), indicating limited focus on genital leptospirosis, and 83.5% lacked clinical information, limiting links between strains and manifestations. Overall, bovine leptospirosis is worldwide distributed, but gaps persist in marker standardization for single-locus sequencing and genome availability. Standardized genotyping, increasing genital sampling, and improving clinical metadata are essential for refining surveillance and clarifying the role of emerging species. Full article

In the context of rapid urbanization, the traditional spatial fabric and cultural connotations of historic districts are increasingly threatened, leading to growing problems such as architectural homogenization and weakened public identity. As an important dimension linking spatial form and public cognition, urban memory has gradually become a key entry point for the study of historic district conservation and renewal. At the same time, the large volume of user-generated content accumulated on social media provides a new data foundation and research pathway for architectural and urban memory studies. Taking the Sajinqiao area in Xi’an as the study area, this study uses Weibo texts containing the keyword “Sajinqiao” from 2018 to 2025 as the basic dataset. A Chinese-RoBERTa pretrained language model was employed to identify and screen high-focus Weibo samples, and a classification framework of five types of memory elements was constructed, including roads, areas, nodes, business units, and food entities. On this basis, memory elements were extracted, standardized, and quantified in terms of memory intensity to analyze their evolutionary characteristics. The results show that, first, urban memory in the Sajinqiao area exhibited marked stage-based fluctuations during the study period. Second, business- and consumption-related elements remained dominant in the type structure over the long term. Third, core urban memory was primarily supported by local food entities and related business units, indicating that public memory gradually shifted from experience-oriented memory to destination-oriented memory. This study provides an operational framework for the identification, quantification, and dynamic assessment of urban memory in historic districts, and offers empirical support for memory-oriented conservation and renewal strategies in the Sajinqiao area and similar historic districts. Full article

The Amazon Basin and its rivers play a vital role in regional biodiversity, the carbon cycle, and socio-economic security. Through erosion and deposition, river planforms change over time, affecting local infrastructure, food security, and changes to ecosystems. Long-term monitoring is essential for observing these dynamics. Synthetic Aperture Radar (SAR) provides a method to consistently map river planform dynamics across large areas because it is largely independent of atmospheric conditions. This study presents an approach for deriving river planform metrics across the entire Amazon Basin using Sentinel-1 C-band SAR data. This approach followed three main steps: water mask generation, validation of the data and river metrics extraction. Sentinel-1 imagery from 2017 to 2025 was composited into quarterly mean images, after which Otsu thresholding was applied to derive water classifications. Additional post-processing steps were applied to reduce terrain- and seasonal effects. The final water masks were divided into water-change classes, validated using stratified sampling and achieved an overall accuracy of 98.5%. Quarterly river planform metrics, including sinuosity, mean channel width and migration rate, were derived using channel centerline extraction, but due to a lack of in situ validation data the river metric values have not been validated. The resulting time series provide insights into how river planform changes across all Amazon sub-basins from 2017 to 2025 can be monitored using SAR-based methods. The results reveal spatial differences in river dynamics between tributaries, mostly depending on flow pattern, up- or downstream path and location in the upper, middle or lower Amazon Basin. These findings demonstrate the potential of SAR time series for monitoring large-scale river planform dynamics. Full article

Indian metropolitan cities such as Mumbai grapple with rapid urbanisation, extreme urban density, high built-up areas, loss of green cover, and shrinking open spaces, resulting in increased impermeable surfaces, urban heat island effects, and frequent flooding occurrences. Modern stormwater management has increasingly been characterised by integrated grey-green approaches; however, cities in the Global North benefit from established policies, technical expertise, and financial resources that enable the systematic and large-scale integration of Blue-Green Infrastructure (BGI) through district-wide geospatial assessment frameworks, unlike many cities in the Global South. Despite growing interest in nature-based stormwater solutions, there remains a dearth of geospatial empirical research from India examining the placement, distribution, performance, and functionality of BGI integrated with existing stormwater management systems in cities such as Mumbai. Furthermore, hydrological modelling using tools such as the Storm Water Management Model (SWMM) for the design, planning, and implementation of BGI in Indian cities remains largely unexplored. This study explores the role of BGI strategies in improving urban stormwater management within high-density Indian cities under a 25-year return period extreme rainfall scenario. Using an integrated approach that combines QGIS-based spatial analysis with EPA-SWMM hydrologic-hydraulic modelling, the research examines runoff behaviour, identifies flooding hotspots, and evaluates the effectiveness of Low Impact Development (LID)-based BGI measures such as permeable pavements, infiltration trenches, and green roofs applied at the ward level in Mumbai’s F/North and G/North Wards. Detailed land use classification, spatial mapping, and rainfall simulation corresponding specifically to a 25-year return period rainfall event was used to assess pre- and post-intervention conditions. The findings indicate that the applied BGI measures led to a 12.6% reduction in peak runoff (137.6 m³/s to 120.2 m³/s) and a 5.5% decrease in total runoff volume (783,510 m³ to 740,410 m³). More importantly, the peak flooding flow rate decreased by 45% (94.1 m³/s to 51.7 m³/s), demonstrating that BGI measures can efficiently reduce peak flooding flows by extending runoff hydrographs during extreme rainfall events. These findings are specifically applicable to the simulated 25-year return period extreme rainfall scenario and may vary under different rainfall intensities or return periods. Less extreme events could potentially experience even greater relative reductions or prevent flooding altogether, while also easing downstream hydraulic loads. Overall, strategically placed BGI interventions can significantly reduce surface runoff and peak flow, thereby enhancing stormwater resilience within spatially constrained urban environments. This study provides a replicable, data-driven framework for catchment-scale stormwater planning in dense Indian cities under extreme rainfall conditions, offering practical insights into methods, local contextual considerations, and spatial planning strategies for policymakers and urban planners seeking to retrofit and adapt existing infrastructure under increasing hydrologic stress and climate variability. Full article

Early non-invasive correction of neonatal auricular deformities is highly dependent on timely and precise diagnosis. However, clinical practice is often compromised by the subjectivity of visual assessments and the lack of objective tracking metrics, which frequently leads to missed optimal treatment windows. To address these challenges, we developed an interpretable deep learning-based diagnostic system for the automated screening and fine-grained classification of these deformities. Methodologically, a large-scale, multi-source dataset (n = 4644) was curated to support model training. The system pairs an automated object detector (YOLOv11) for background-reduced region-of-interest isolation with a cascaded classification pipeline optimized via ConvNeXt-Tiny. Crucially, we introduced a supervised contrastive learning module to project high-dimensional morphological features into a continuous severity score, enabling quantitative longitudinal tracking of therapeutic efficacy. To evaluate generalization and robustness, the framework underwent rigorous evaluation across three independent real-world cohorts and one controlled synthetic stress test. The system achieved 88.2% accuracy (Area Under the Curve (AUC): 0.949) in binary screening and 87.4% accuracy (macro-AUC: 0.976) in multi-class subtyping on the internal baseline. To enhance interpretability and build clinical trust, Gradient-weighted Class Activation Mapping (Grad-CAM) was utilized to explore the spatial distribution of the model’s attention, which frequently aligned with key anatomical landmarks. Furthermore, the learned severity scores robustly quantified post-intervention improvements (p = 0.0004), effectively capturing subtle anatomical normalization. While validation for rare subtypes remains underpowered, and the severity score currently functions mainly as a learned morphological similarity index requiring future clinical calibration, this study ultimately provides an objective and standardized web-based tool to facilitate the early intervention and precision management of neonatal auricular anomalies. Full article

The digitization of existing building stock often depends on legacy 2D raster floor plans (scanned drawings, PDF exports, or photographs) because structured building information models are frequently unavailable for older properties. Manual measurement and visual inspection of such documents are time consuming and error prone. This paper presents an integrated deep learning pipeline that extracts semantic information from unstructured two-dimensional floor plan images of existing structures and supports preliminary compliance screening via locally deployed large language models. The pipeline employs YOLOv8 for the localization and classification of 18 architectural symbols and furniture items, and a U-Net with a ResNet34 encoder for the semantic segmentation of walls and interior room spaces. To translate pixel-level predictions into physical metrics, we implement an area calculation module based on user-defined reference scale calibration. An LLM evaluation module, deployed locally via Ollama with a retrieval-augmented generation pipeline, interprets extracted room metrics and flags potential non-compliance against referenced residential design guidelines; it is intended for the assessment of existing layouts rather than generative co-design. We expand a core dataset of 101 manually annotated source floor plans to 303 augmented instances using label-aligned geometric transformations, while reporting generalization in terms of the 101 unique source plans. On the held-out validation split (10 source plans), YOLOv8 achieves 92.3% mAP50 versus 87.2% for a Faster R-CNN reference model on the same data split (detection baselines differ in training epochs and pretraining; see Experiments); U-Net achieves 95.71% mIoU, surpassing DeepLabv3+ (93.2%) under matched segmentation training settings. The system is deployed as an interactive web application for legacy building survey and preliminary regulatory review when only two-dimensional documentation is available. Full article

The growing complexity of web attacks highlights the need for adaptive, intelligent defense systems that overcome the limitations of traditional rule-based web security. Thus, the architecture proposed in this paper integrates data-driven deep learning with deterministic rule-based logic to enhance real-time detection accuracy and adaptability in dynamic web threat environments. The practical integration of a deep learning-based Gated Recurrent Unit (GRU) model with ModSecurity, an open-source Web Application Firewall (WAF), is employed to improve the detection and classification of malicious HTTP requests. The model, pre-trained on a large labeled up-to-date dataset of web traffic and attack types collected post-2020, is designed to classify requests in real-time, identifying both whether a request is malicious and the corresponding attack category (e.g., SQL Injection, Cross-Site Scripting, Command Injection). We demonstrate how the trained model is incorporated into ModSecurity’s inspection pipeline, allowing it to analyze real-time web traffic alongside traditional rule-based inspection. This hybrid approach aims to significantly reduce false positives and improve adaptability to new attack patterns. Evaluation metrics such as accuracy, receiver operating characteristic (ROC), area under the curve (AUC), Principal Component Analysis (PCA), confusion matrix, and t-Distributed Stochastic Neighbor Embedding (t-SNE) visualization are discussed, along with performance considerations and implementation architecture. The integration presents a robust framework for ML-improved intelligent web security defense. Full article

Objectives: Multimodal large language models (MLLMs) have shown potential for medical image classification. We evaluated four optimization strategies in two MLLMs—GPT-4o (gpt-4o-2024-08-06) and Gemini 2.5 Flash-Lite—for ultrasound-based thyroid nodule malignancy classification using two public datasets and a clinical cohort of nodules with atypia of undetermined significance (AUS) cytology. Methods: Text prompting, few-shot learning, fine-tuning, and a hybrid strategy combining fine-tuning with few-shot learning were evaluated for each model. Performance was assessed using the Digital Database of Thyroid Images (DDTI; n = 80), a 1000-image test subset of TN5000, and an institutional AUS cohort with surgical pathology (n = 84). In the AUS cohort, the best-performing strategy was compared with the consensus classification of three endocrinologists and the American Thyroid Association (ATA) ultrasound risk stratification. Results: For GPT-4o, the hybrid strategy achieved the highest area under the receiver operating characteristic curve (AUC) in DDTI (0.866), TN5000 (0.689), and the AUS cohort (0.836). In the AUS cohort, its specificity was higher than that of endocrinologist consensus and ATA risk stratification when only high-suspicion nodules were classified as malignant (95.1% vs. 70.7% and 70.7%; p = 0.002 and p = 0.001, respectively), while sensitivity did not differ significantly (72.1% vs. 74.4% and 79.1%, respectively; both p > 0.05). However, the hybrid model misclassified 12 of 43 malignant nodules, corresponding to a false-negative rate of 27.9%. When high- and intermediate-suspicion ATA categories were classified as malignant, ATA sensitivity increased to 83.7% and specificity decreased to 56.1%; the hybrid model had a higher AUC than ATA risk stratification (0.836 vs. 0.749; p = 0.017). For Gemini 2.5 Flash-Lite, few-shot learning, fine-tuning, and the hybrid strategy did not improve AUC relative to text prompting in any dataset. Conclusions: The hybrid strategy produced the most consistent performance gains for GPT-4o across the three datasets but did not improve Gemini 2.5 Flash-Lite. The optimized GPT-4o model achieved high specificity in the diagnostically challenging AUS cohort, although its false-negative rate limits its use as a stand-alone diagnostic tool. Further validation in larger, prospective multicenter cohorts is required before clinical use. Full article

Background/Objectives: Given rising lung cancer incidence and limited data on pulmonologist-performed ultrasound-guided transthoracic core biopsy (US-TTCB), in this study, we evaluated diagnostic yield and safety for pleural or pulmonary lung masses, using Clavien–Dindo classification to standardize complication reporting. Methods: We retrospectively reviewed single-center pulmonologist-performed US-TTCB using a MEDONE biopsy gun with a 16 G/18 G Tru-Cut needle between January 2019 and December 2025. The primary endpoints were diagnostic yield, defined as specific malignant or benign histology, and complication rate. Non-diagnostic results were assessed using available clinical/imaging follow-up. Univariate analyses screened candidate correlates, and a prespecified computer tomography (CT)-completed subanalysis (n = 67) used multivariable logistic regression and receiver operating characteristic (ROC) analysis to assess CT lesion size discrimination. Results: Diagnostic yield was 84.2% (202/240); complications occurred in 12.1% (29/240), including one Clavien–Dindo Grade III event (0.4%). In the CT-completed subset (n = 67), diagnostic yield was independently associated with CT lesion size (aOR 1.03/mm, 95% CI 1.00–1.05; p = 0.022) and Chronic Obstructive Pulmonary Disease (COPD) (aOR 2.30, 95% CI 1.06–4.96; p = 0.034); CT lesion size showed an area under the curve (AUC) of 0.717 for predicting yield. Diagnostic yield remained stable over time (84.2% in first vs. second half; p = 1.00), with no association between case order and yield (OR 0.999; p = 0.64). Conclusions: US-TTCB of pleural/pulmonary masses achieved a high diagnostic yield with minimal major complications. Large CT dimension and COPD were associated with higher diagnostic success, and CT size provided fair discrimination for predicting yield; findings should be interpreted in the context of the retrospective single-center design and the restricted CT-completed subset. Full article